A critical analysis and interpretation of the data published for complexation (conditional stability constants and complexation capacities) of trace metals (copper, Cu(II); zinc, Zn(II); lead, Pb(II); and cadmium, Cd(II)) in whole seawater and freshwater samples over the past 34 yr is presented. We propose that it is not necessary to invoke the presence of any particular classes of ligands (e.g., the so-called L1 and L2 ligands) to explain the degree of metal complexation observed. Consideration of the presence of a range of many different types of ligands could explain the observed linear relationship between the stability of complexes (log equation image) and the effective metal ion loading (complexation capacity, equation image), without the need to assume the presence of any stable specific complexants. Apparently stronger binding sites are utilized at lower metal ion loadings; progressively weaker sites contribute to complexation at higher metal ion concentrations. Taking into account the detection window of the technique employed greatly improves the internal consistency of the overall data. When comparable windows are used, no significant difference is observed between the log equation image relationship for Cu(II) and that for Zn(II), Pb(II), or Cd(II). In natural waters, the complexing material present will be a combination of that derived from terrestrial sources and that formed in situ in the water column. An important observation arising from our analysis is that the binding curves for natural ligands in whole natural-water samples lie between the binding curves for compounds representative of each these sources, viz. isolated aquatic fulvic acids and biota.